1 Programs and scripts for automated testing of Zstandard
2 =======================================================
4 This directory contains the following programs and scripts:
5 - `datagen` : Synthetic and parametrable data generator, for tests
6 - `fullbench` : Precisely measure speed for each zstd inner functions
7 - `fuzzer` : Test tool, to check zstd integrity on target platform
8 - `paramgrill` : parameter tester for zstd
9 - `test-zstd-speed.py` : script for testing zstd speed difference between commits
10 - `test-zstd-versions.py` : compatibility test between zstd versions stored on Github (v0.1+)
11 - `zstreamtest` : Fuzzer test tool for zstd streaming API
12 - `legacy` : Test tool to test decoding of legacy zstd frames
13 - `decodecorpus` : Tool to generate valid Zstandard frames, for verifying decoder implementations
16 #### `test-zstd-versions.py` - script for testing zstd interoperability between versions
18 This script creates `versionsTest` directory to which zstd repository is cloned.
19 Then all tagged (released) versions of zstd are compiled.
20 In the following step interoperability between zstd versions is checked.
22 #### `automated-benchmarking.py` - script for benchmarking zstd prs to dev
24 This script benchmarks facebook:dev and changes from pull requests made to zstd and compares
25 them against facebook:dev to detect regressions. This script currently runs on a dedicated
26 desktop machine for every pull request that is made to the zstd repo but can also
27 be run on any machine via the command line interface.
29 There are three modes of usage for this script: fastmode will just run a minimal single
30 build comparison (between facebook:dev and facebook:release), onetime will pull all the current
31 pull requests from the zstd repo and compare facebook:dev to all of them once, continuous
32 will continuously get pull requests from the zstd repo and run benchmarks against facebook:dev.
35 Example usage: python automated_benchmarking.py
39 usage: automated_benchmarking.py [-h] [--directory DIRECTORY]
40 [--levels LEVELS] [--iterations ITERATIONS]
41 [--emails EMAILS] [--frequency FREQUENCY]
42 [--mode MODE] [--dict DICT]
45 -h, --help show this help message and exit
47 directory with files to benchmark
48 --levels LEVELS levels to test e.g. ('1,2,3')
49 --iterations ITERATIONS
50 number of benchmark iterations to run
51 --emails EMAILS email addresses of people who will be alerted upon
52 regression. Only for continuous mode
54 specifies the number of seconds to wait before each
55 successive check for new PRs in continuous mode
56 --mode MODE 'fastmode', 'onetime', 'current', or 'continuous' (see
57 README.md for details)
58 --dict DICT filename of dictionary to use (when set, this
59 dictionary will be used to compress the files provided
63 #### `test-zstd-speed.py` - script for testing zstd speed difference between commits
67 This script creates `speedTest` directory to which zstd repository is cloned.
68 Then it compiles all branches of zstd and performs a speed benchmark for a given list of files (the `testFileNames` parameter).
69 After `sleepTime` (an optional parameter, default 300 seconds) seconds the script checks repository for new commits.
70 If a new commit is found it is compiled and a speed benchmark for this commit is performed.
71 The results of the speed benchmark are compared to the previous results.
72 If compression or decompression speed for one of zstd levels is lower than `lowerLimit` (an optional parameter, default 0.98) the speed benchmark is restarted.
73 If second results are also lower than `lowerLimit` the warning e-mail is sent to recipients from the list (the `emails` parameter).
76 - To be sure that speed results are accurate the script should be run on a "stable" target system with no other jobs running in parallel
77 - Using the script with virtual machines can lead to large variations of speed results
78 - The speed benchmark is not performed until computers' load average is lower than `maxLoadAvg` (an optional parameter, default 0.75)
79 - The script sends e-mails using `mutt`; if `mutt` is not available it sends e-mails without attachments using `mail`; if both are not available it only prints a warning
82 The example usage with two test files, one e-mail address, and with an additional message:
84 ./test-zstd-speed.py "silesia.tar calgary.tar" "email@gmail.com" --message "tested on my laptop" --sleepTime 60
87 To run the script in background please use:
89 nohup ./test-zstd-speed.py testFileNames emails &
92 The full list of parameters:
95 testFileNames file names list for speed benchmark
96 emails list of e-mail addresses to send warnings
99 -h, --help show this help message and exit
100 --message MESSAGE attach an additional message to e-mail
101 --lowerLimit LOWERLIMIT
102 send email if speed is lower than given limit
103 --maxLoadAvg MAXLOADAVG
104 maximum load average to start testing
105 --lastCLevel LASTCLEVEL
106 last compression level for testing
107 --sleepTime SLEEPTIME
108 frequency of repository checking in seconds
111 #### `decodecorpus` - tool to generate Zstandard frames for decoder testing
112 Command line tool to generate test .zst files.
114 This tool will generate .zst files with checksums,
115 as well as optionally output the corresponding correct uncompressed data for
120 ./decodecorpus -ptestfiles -otestfiles -n10000 -s5
122 will generate 10,000 sample .zst files using a seed of 5 in the `testfiles` directory,
123 with the zstd checksum field set,
124 as well as the 10,000 original files for more detailed comparison of decompression results.
127 ./decodecorpus -t -T1mn
129 will choose a random seed, and for 1 minute,
130 generate random test frames and ensure that the
131 zstd library correctly decompresses them in both simple and streaming modes.
133 #### `paramgrill` - tool for generating compression table parameters and optimizing parameters on file given constraints
135 Full list of arguments
137 -T# : set level 1 speed objective
138 -B# : cut input into blocks of size # (default : single block)
139 -S : benchmarks a single run (example command: -Sl3w10h12)
148 --zstd= : Single run, parameter selection syntax same as zstdcli with more parameters
149 (Added forceAttachDictionary / fadt)
150 When invoked with --optimize, this represents the sample to exceed.
151 --optimize= : find parameters to maximize compression ratio given parameters
152 Can use all --zstd= commands to constrain the type of solution found in addition to the following constraints
153 cSpeed= : Minimum compression speed
154 dSpeed= : Minimum decompression speed
155 cMem= : Maximum compression memory
156 lvl= : Searches for solutions which are strictly better than that compression lvl in ratio and cSpeed,
157 stc= : When invoked with lvl=, represents percentage slack in ratio/cSpeed allowed for a solution to be considered (Default 100%)
158 : In normal operation, represents percentage slack in choosing viable starting strategy selection in choosing the default parameters
159 (Lower value will begin with stronger strategies) (Default 90%)
160 speedRatio= (accepts decimals)
161 : determines value of gains in speed vs gains in ratio
162 when determining overall winner (default 5 (1% ratio = 5% speed)).
163 tries= : Maximum number of random restarts on a single strategy before switching (Default 5)
164 Higher values will make optimizer run longer, more chances to find better solution.
165 memLog : Limits the log of the size of each memotable (1 per strategy). Will use hash tables when state space is larger than max size.
166 Setting memLog = 0 turns off memoization
167 --display= : specify which parameters are included in the output
168 can use all --zstd parameter names and 'cParams' as a shorthand for all parameters used in ZSTD_compressionParameters
169 (Default: display all params available)
170 -P# : generated sample compressibility (when no file is provided)
171 -t# : Caps runtime of operation in seconds (default: 99999 seconds (about 27 hours))
172 -v : Prints Benchmarking output
173 -D : Next argument dictionary file
174 -s : Benchmark all files separately
175 -q : Quiet, repeat for more quiet
176 -q Prints parameters + results whenever a new best is found
177 -qq Only prints parameters whenever a new best is found, prints final parameters + results
178 -qqq Only print final parameters + results
179 -qqqq Only prints final parameter set in the form --zstd=
180 -v : Verbose, cancels quiet, repeat for more volume
181 -v Prints all candidate parameters and results
184 Any inputs afterwards are treated as files to benchmark.